The proposed research will perform secondary data analyses on 10 waves of data from the Cardiovascular Health Study (CHS). The CHS includes 5,888 community-dwelling older adults sampled from several US regions. At baseline, CHS participants had magnetic resonance imagining scans and other measures. Those with dementia were screened from the follow-up sample for dementia, but were retested on other measures. After 10 years, the dementia-free participants were retested and 480 cases of incident dementia and 577 of Mild Cognitive Impairment (MCI) were observed. We will use this sample to meet our primary goal which is to examine incident life events as risk factors for cognitive decline, MCI, and incident dementia. Our specific aims are to use theoretical models of stress to test hypotheses about direct relationships of four incident stressful life events (SLEs) (i.e., death of a loved one, caregiving, illness of a loved one, poor relationship with loved one) with cognitive decline/incident dementia; and indirect relationships of SLEs, mediators, and cognitive decline/dementia. The mediators have theoretical and empirical support. They include psychosocial (social supports, quality of life, depression, sleep problems), behavioral (drinking, smoking, exercise, diet), and physiological (metabolic syndrome, inflammation) measures, as well as incident illnesses (hypertension, diabetes, etc). We will also examine moderators, namely interactions of SLEs with putative etiologic factors (the above cited mediators) and with genetic (APOE), physical (illnesses) and chemical (medications). Because these measures were repeated several times over the course of 10 years, the CHS allows doubly- prospective analyses. Such analyses are rare in the SLE literature. They provide important information about individuals before they experience a SLE and before they experience cognitive decline and/or incident dementia. For example, if psychophysiological measures mediate relationships of incident SLEs with cognitive decline/dementia, we will be able to estimate whether this was influenced by the SLE itself, the putative mediators before the SLE, or the interaction of the two. We will also be able to assess whether some SLE are acute or chronic (e.g. caregiving) and whether chronic experiences interact with acute events to exacerbate the influence of each other on the outcomes. These relationships could have major implications for the design of programs for persons exposed to potentially harmful stressors. In the next 20 years, the US population over 65 years old will grow much faster than the proportion that is less than 35 years old. Given the relationship between age and dementia, government agencies have warned of a dementia epidemic. Dementia causes suffering for victims and families and is very costly to health care. To mitigate this crisis, we must understand the risks for developing cognitive problems/dementia. Genetic-environmental risks have been identified for dementia, but limited attention has been given to SLEs. This is surprising because SLEs and dementia are associated with cognitive, psychophysiological, and behavioral problems.